A Lega and English dictionary comp. by Robert Botne (review)

672 LANGUAGE, VOLUME 74, NUMBER 3 (1998) Biloa's doctoral dissertation, Tuki has become the most extensively analyzed Bantu language for the treated topics within the P&P theory of UG, as is evident from B's copious discussions of languages previously analyzed in the same framework, mainly European and Asian, but total absence of references to Bantu studies. It is fairer to credit the publisher for energetically trying to fill large gaps in the linguistic record on obviously limited financial resources than to bemoan the appalling proofreading and editing of its publications. Nevertheless, this volume will offend the squeamish, averaging over three misprints per page, sometimes with serious consequences, and overeconomizing on space by splitting tables from their headings (e.g. 12-13) and examples from their morpheme glosses (e.g. 68-69) instead of changing a page break. Considering how densely argued and richly exampled the book is, the editing is not all that bad. Among the topics extensively discussed are pronominal reference (including in relation to quantifiers ), clefting, relativization, matrix and embedded WH-questions, sentential complements, and verb focus. Characteristic of P&P theory, concern involves problems ofdeciding betweenbase generation and movement from what node to what node. Readers need not accept the theory to profit from the rich syntactic paradigms invoked for the analyses. B skillfully locates many issues where Tuki data allow him to offer revisions to prior P&P theory, e.g. the independence of subject postposing, illicit in Tuki, from its obligatory verbal subject-marking (34, 95). He displays less knowledge of Bantu syntactic diversity or recognition of issues relevant to the discussed syntactic processes unless they are already established in the P&P literature. Thus, in arguing that the verb's object marker is not a 'pronoun' but an 'agreement' marker to an overt or covert postverbal pronoun, he criticizes unnamed ' ' 'traditional bantuists " ' [sic] (171), ignoring the many documented Bantu languages in which an object marker is incompatible with a postverbal object. He similarly neglects discussing whether either or both of two postverbal objects can freely condition such agreement , e.g. in 'he minded the children for me' (i.e. . . . mind-for me them). A minimal general grammatical sketch of Tuki in Ch. 1 is of very uneven quality, often inaccurate or misleading in view of discussion or examples in later chapters. For example, Tuki's typical Bantu independent pronouns are introduced as 'subject pronouns' (31) but later revealed to also have uses as objects of verbs and prepositions (173). Similarly, B introduces PO 'present' as a tense referring to 'actions occurring today' (12), but later examples show it can co-occur with nambari 'tomorrow' (167). Occasional inaccuracies or oversimplifications also occur in the more technical chapters. Particularly shocking is an example inadvertently indicating prepositionstranding (68), a process unknown elsewhere in Bantu. Indeed, B later indicates that prepositionstranding is impossible in Tuki (152), and it seems that the earlier example contained an oversimplified gloss, na 'to' o when the primary pitch accent, at the time of the innovation, shifted to another syllable (139). Why would change ofpitch lead to backing and rounding? L's (132) uncritical citing of Hirt's use of Russian e ...